Open Collections

Verbal source code descriptor

University of British Columbia

Program comprehension and understanding is a demanding cognitive task in software maintenance. Program understanding involves the construction of a mental model of the software. Researchers introduced various cognitive models for program understanding, such as the top-down model [8], bottom-up model [7], the knowledge-based model [9], the systematic and as-needed program understanding model [11], and the integrated program understanding meta-model [12]. Many tools have been developed to assist program comprehension based on aspects of the above models. These tools basically create visual high-level information about the software under investigation, particularly at the structure and design level, and are designed to form the mental models of the software in hand. This research presents a new approach to program comprehension with the goal of reducing the cost of existing tools and overcoming some of their limitations, such as language dependency, and the insufficiency of providing information for complete program understanding. Program understanding information basically exists at the time of programming, and can be captured with less cost and effort; however some problems, such as mental loading problem, prevent programmers from providing such information in the form of written documentation. I have developed a verbal documentation strategy to reduce the overhead of typing at the time of programming. The hypothesis is that we can capture more useful understanding information from the programmer using this strategy. . The Verbal Source Code Descriptor (VSCD) System is introduced as a tool to help programmers document their source code verbally, while writing the code. A preliminary evaluation with 12 programmers was carried out to evaluate the feasibility and ease of using the VSCD System. This is compared with more conventional implementation documenting techniques, which is still in the form of putting comments inside the source code. The results indicate the VSCD system may be useful and efficient in generating comments. Future studies must verify the quality of VSCD comments and investigate the efficiencies of using this strategy to make program comprehension more flexible and faster. These studies may combine the VSCD strategy with existing methods to investigate the effects of this approach.

Thesis/Dissertation

application/pdf

2009-10-17

For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.

http://hdl.handle.net/2429/13954

Electrical and Computer Engineering

University of British Columbia

UBCV

DSpace